Long-term hydrogeophysical monitoring of the internal conditions of river levees

To evaluate the vulnerability of the earthen levee of an irrigation canal in San Giacomo delle Segnate, Italy, a customized electrical resistivity tomography (ERT) monitoring system was installed in September 2015 and has been continuously operating since then. Thanks to a meteorological station deployed at the study site, we could investigate the relationship between the inverted resistivity values and different parameters, namely air temperature, rainfall and water level in the canal. Air temperature seems to have a minor but not negligible influence on resistivity variations, especially at shallow depth. A model of soil temperature versus depth was used to correct resistivity sections for air temperature variations through the different seasons. Changes of the water level in the canal and rainfall significantly affect measured resistivity values. At the study site, the most important variations of resistivity are related to saturation and dewatering processes in the irrigation periods. Although we explored the effect of drawdown procedures on resistivity data, this process, causing rapid variations of resistivity values, is still not completely understood because the canal is rapidly emptied during rainfall events. Therefore, the effect of variations of the water level in the canal on levee resistivity cannot be distinguished from the effect of rainfalls. To study the effect of water level variations alone, we considered the beginning of the irrigation period when the dry canal is gradually filled and we observed a smooth trend of resistivity changes. The effect of rainfall on the data was studied during different periods of the year and at different depths of the levee so that the resistivity variations could be evaluated under different conditions. To convert the inverted resistivity sections into water content maps, an empirical and site-dependent relationship between resistivity and water content was obtained using core samples. Water content data can then be used for the implementation of stability analysis using custom modeling. This study introduces an efficient technique to monitor earthen levees and to control the evolution of seepage and water saturation in pseudo-real time. Such a technique can be exploited by Public Administrations to reduce hydrogeological risks significantly.